Faculty & Research

The Development of The Personality Audit: A Psychodynamic Multiple Feedback

Assessment Instrument by

M. Kets deVries P. Vrignaud K. Korotov E. Engellau

and E. Florent-Treacy

2004/71/ENT

Working Paper Series

The Development of The Personality Audit:

A Psychodynamic Multiple Feedback Assessment

Instrument

Manfred F.R. Kets de Vries

Raoul de Vitry d’Avaucourt Clinical Professor of Leadership Development,

Director of INSEAD Global Leadership Center

INSEAD, Fontainebleau (France) and Singapore

Manfred.Kets-de-Vries@insead.edu

Pierre Vrignaud

Professor Psychologie du développement. Université Paris X. Laboratoire de

Psychologie du développement social et émotionnel. EA 1588.

France

Vrignaud.pierre@wanadoo.fr

Konstantin Korotov

PhD Candidate, INSEAD

Konstantin.Korotov@insead.edu

Elisabet Engellau

Adjunct Clinical Professor of Management, INSEAD

Elisabet.Engellau@insead.edu

Elizabeth Florent-Treacy

Research Project Manager, INSEAD

Elizabeth.Florent@insead.edu

1

Abstract

The objective of this working paper is to describe the design and psychometric

properties of The Personality Audit (PA), a psychodynamic multiple feedback

instrument used to help participants assess seven personality dimensions important in

human behavior. The instrument, in contrast with other tools that can be used to

clarify the inner theater of individuals, is designed not only to report information

given by the test-taker but also to reflect the perceptions of observers representing

both the test-taker’s public and private spheres. Given the serious gap that often exists

between an individual’s self-perception and the perceptions others have of him/her, a

multiple feedback approach presents a more concise and accurate picture of behavior

than more traditional personality tests, providing information helpful to executives

aspiring to become effective leaders.

This working paper describes in detail the conceptual foundations of the

questionnaire, and the psychometric methods used to confirm its validity and

reliability. In addition, it discusses possible directions for future research.

Keywords: 360-degree instrument; human development; life cycle; motivational

need systems; psychodynamic approach; personality assessment; executive

functioning; inner theater

2

So was it when my life began

So is it now I am a man

So be it when I shall grow old

Or let me die!

The Child is the father of the Man

—William Wordsworth

The heart has its reasons, which reason knows not.

—Blaise Pascal

Personality Assessment for Executives

Above the entrance to the temple of Apollo in Delphi was written “Know Thyself.”

This famous inscription remains as valid today as it was thousands of years ago. To be

effective as leaders, executives have to start with themselves. They have to understand

the reasons for doing what they do. They have to study their own human motivation to

truly understand their decisions and actions. A psychodynamic and psychosocial

approach to the study of personality accomplishes exactly this task, rendering the

unsaid said and the unconscious conscious. Studying people from this perspective

means taking into account their relational world, paying attention to the forces of

human development, and considering the dynamics of emotional management. Such

an approach creates a richer, more three-dimensional appreciation of human behavior.

Personality assessment cannot take place on a desert island (whether real or

metaphorical). The very concept of assessment implies a comparison of one’s

personality functioning with that of others. Such comparison can give test-takers

useful insights into their personality characteristics, strengths and weaknesses, and

developmental needs. The Personality Audit discussed in this working paper is a

relatively simple tool designed to provide an assessment of seven personality

dimensions important in human functioning. Dimensions such as

introversion/extroversion, distrustful/trustful, and others reveal the psychodynamic

and psychosocial forces that drive any given test-taker. An interpretation of these

dimensions provides test-takers with a deeper understanding of their own behavior.

3

By providing insight into the extent to which conscious observable and unconscious

invisible processes influence behavior in individuals, dyads, groups, and

organizations, The Personality Audit supports a better understanding of interpersonal

relationships, recurring conflict patterns, and the meaning of one’s actions and

experiences. The instrument is designed to report information given by the test-taker

as well as reflect the perceptions of at least three “observers,” representing both the

private and public spheres.

Although the observations of all the various “others” can be given anonymously—and

indeed the literature on multiple feedback systems often supports that approach (e.g.,

Carlson, 1998)—we strongly recommend that the observers’ names remain attached

to their scores. This helps the test-taker recognize and understand differences in

perceptions, and it allows the test-taker to later engage in meaningful discussion with

those who provided feedback. As an added benefit, keeping the observers’ names

attached to the scores may increase the observers’ accountability for helping the test-

taker take action on the basis of the results obtained (London, Smither, and Adsit,

1997).

In order for test results to reflect potential differences in upward and downward styles

of leadership, we recommend that at least one superior and one subordinate be

included among those asked to give feedback. In addition, because individuals may be

perceived differently in public life (at work) than in private life (at home), we stress

the value of having both work-related observers and observers who usually see the

test-taker in his/her non-work life. Of particular interest is the potential divergence in

the perception of behavior between the test-taker and the spouse or significant other.

To the best of our knowledge, no other instrument, including existing managerial 360-

degree feedback questionnaires (e.g., Kouzes and Posner, 2001) and personality tests,

allows for incorporation of observers from both public and private spheres.

Because the instrument incorporates feedback from spouse/partner, friends, and

subordinates and superiors at work, the PA helps test-takers identify “blind spots” in

personal functioning. The resulting insights can then be used as the basis for the

formulation of appropriate development plans. The instrument is helpful to executives

4

involved in leadership development programs, executives involved in coaching and

counseling related to growth and development, and even couples involved in marital

counseling.

Rationale

The Personality Audit grew out of the need for a relatively simple instrument that

could clarify the various motivational characteristics of executives. There are a

number of assessment tools that try to accomplish the task of looking into the test-

taker’s personality, but they are relatively difficult to use. Projective tests such as the

Thematic Apperception Test (Murray, 1938; Shneidman, Joel, & Little, 1951) and the

Rorschach Test (Goldfried, Stricker, & Weiner, 1971) provide valuable insights into

the inner world of an individual, but their application is extremely complicated.

Making sense out of inkblots and interpreting stories on the basis of presented pictures

can be cumbersome, and the administration and interpretation of such tests require

considerable training. Furthermore, the unstructured nature of the test stimuli in these

projective tests creates problems of empirical reliability and validity, as well as

consistency of interpretation.

Some of the same observations apply to well-known comprehensive pencil-and-paper

tests of personality, such as the Minnesota Multiphasic Personality Inventory

(Dahlstrom, Welsh, & Dahlstrom, 1972; Hathaway & McKinley, 1943), the California

Personality Inventory (Gouch, 1975), the 16 Personality Factor Questionnaire (Cattell,

1957; Cattell, Eber, & Tatsuoka, 1970), Millon’s Clinical Multiaxial Inventory

(Millon, 1983), and Hogan’s Personality Inventory (Hogan & Hogan, 1992). Again,

these instruments can be extremely helpful in providing insights into personality

functioning; however, like the Rorschach and Thematic Apperception tests, they are

relatively difficult to administer and interpret, because they cast an extremely wide net

in their efforts to obtain information about personality. In addition, some of these tests

emphasize major personality pathology.

Tests that were specifically designed to study people with different pathologies have

limited usefulness when applied to a non-clinical population. Furthermore, some of

the available tests were designed to probe surface behavior only. Such tests provide

no insights into the test-taker’s inner world. Moreover, there are currently no

5

instruments that look into those aspects of the inner theater of individuals that are

particularly important for the functioning of executives in organizations. Most

importantly, however, all of the available personality instruments deal only with self-

assessment. In contrast, the PA is designed to enable the test-taker to compare his/her

self-perceptions with the observations of others.

Experience has taught us that there is usually a serious gap between what people say

they do and what they actually do—that is, between their self-perception and the

(theoretically more objective) perception that other people have of them. Given the

prevalence of this gap, we saw a multiple feedback approach as necessary to this

instrument’s design. Literature on 360-degree feedback in organizations (Church &

Bracken, 1997; London & Beatty, 1993; Yammarino & Atwater, 1993, 1997; Carlson,

1998; Bland, Edwards, & Kuhi, 1994) agrees that the multiple feedback approach

gives managers in organizations a more accurate view of themselves than self-

evaluation does.

The multiple feedback approach with input from various observers minimizes the so-

called social desirability factor, which reflects the human tendency to present oneself

(consciously or unconsciously) in a more positive light, distorting one’s responses to

achieve that effect. At the same time, inclusion of the self-rating, rather than exclusive

reliance on the arguably more accurate scores provided by observers, not only allows

for comparison with and discussion about the observers’ views, but also sets the stage

for greater acceptance of other people’s views (Mohrman, West, and Lawler, 1989).

Given a target audience of executives, we strove to construct a diagnostic instrument

that was simple but conceptually sound, was aimed at a “normal” population, had a

psychodynamic focus, used the multiple feedback approach (including feedback from

both public and private sides of the test-taker’s life), and could be used as the basis for

a meaningful and helpful discussion about individual behavior. Our challenge was to

develop an instrument that would be easy to complete, easy to understand and

interpret, and easy to translate into action aimed at improvement.

The Personality Audit was developed as a dimensional model of salient personality

traits based on findings from research on human development. A dimensional model

recognizes that the behavior of well-functioning and poorly functioning individuals

6

can be represented as points on a continuum. In other words, such a model portrays an

individual’s personality by indicating quantitative gradations of intensity as positions

on a scale, rather than by offering qualitative, discrete representations of personality.

Conceptual Background

Biological underpinnings of personality

Biologists and psychologists agree that individuals are a product of nature and

nurture. Every developmental line follows an innate timetable. The nature/nurture

interaction involves genetically determined maturational sequences and inherent

potentialities, environmental influences, and personal experiences. The unfolding

sequence of developmental stages and the readiness for successive maturational levels

is biologically determined and innate to each species; but the nature of the parent-

child interaction determines how the child (and later the adult) will interpret and

integrate various experiences. As people experience the world around them, with the

body and intellect that nature endowed, development occurs in several areas:

• Biosocial processes. The brain and other parts of the body adapt and grow as the

environment influences them.

• Cognitive processes. Thought processes, perceptual abilities, and language mastery

all respond to the environment.

• Psychosocial processes. Emotions, personality, and patterns of interpersonal

relationships develop and solidify as the person interacts with the world.

All of these developments—outcomes of the interaction between inherited

predispositions and the environment—result in specific behavior patterns.

To understand the progressive appearance, change, and organization of mental

processes and functions, psychologists refer to a number of developmental schemas

that have been devised to help clarify what makes people the way they are. Three

major conceptual schemas dealing with human development form the foundation on

which the PA is based. We superimposed these three theoretical frameworks on each

7

other during the development of the instrument, to give us a rich description of

personality:

• The work of Erik Erikson (1963, 1968) on the human life cycle, which we look at in

some detail below, was extremely influential.

• The contributions of Joseph Lichtenberg (1989) on motivational need systems,

based on infant observation studies, contributed to the conceptual framework.

• Finally, the work of Silvan Tomkins (Tomkins, 1962, 1963, 1991, 1992) on affect

added the emotional dimension to this study of personality.

Contributions of other psychologists were also used when appropriate. For example,

one of the earliest known developmental schemas was introduced by Sigmund Freud

(1905). Freud conceptualized the existence of oral, anal, and Oedipal stages in the

course of childhood development, each stage based on those bodily areas that serve as

the major source of satisfaction at a particular point in time. This influential view of

human development was further extrapolated by a large number of psychologists,

including Wilhelm Reich, whose study Character Analysis (Reich, 1933), with its

“drive theory” approach to personality diagnosis, was a turning point.

While Freud and Reich focused specifically on body functions, subsequent

psychologists recognized the importance of interpersonal relations during the course

of development (Balint, Ornstein, & Balint, 1972; Greenberg & Mitchell, 1983;

Sullivan, 1953;Winnicott, 1975). Other researchers constructed stages of development

that were dependent on the content and constancy of mental representations. In their

conceptual schemas, they clarified how mental imagery, with its accompanying

emotional experience, increases in sophistication as time passes, and they qualified

the outcome as dependent on the intensity of the interchange with the primary

caretaker (Bowlby, 1969; Emde, 1981; Fairbairn, 1952; Guntrip, 1968; Mahler, Pine,

& Bergman, 1975; Piaget, 1952; Pine, 1985; Spitz, 1965; Stassen Berger, 1998;

Sullivan, 1953).

Erikson’s Stages of Human Development

The work of Erik Erikson is a milestone in the understanding of the different stages in

human development (Erikson, 1963, 1968). Erikson, affirming the nature/nurture

8

linkage, highlights the interface between the biological aspects of psychosexual

development and the cultural and interpersonal needs of each developing individual.

Erikson describes the human life cycle in a number of stages expressed as polarities,

specifying developmental milestones and tasks to be achieved at each stage. These

emerging behavioral modes are viewed as epigenetic, meaning that if any given mode

is not completed at the appropriate developmental phase, that delay will affect future

development. These modes do not disappear even if development continues “on

schedule,” however; an echo of each can be heard in later developmental phases.

Furthermore, each life stage requires a reformulation and resolution of crises from

earlier stages of life. As a result, changes in maturation rarely form a smooth process;

rather, they proceed in leaps and bounds from one stage of development to another.

Because each of Erikson’s stages (or “eight ages of man”) concerns a struggle to

become involved with others, how those “others” respond is vital. In the resolution of

the series of challenges that each developing infant faces, the role of the immediate

environment is of crucial importance. If primary caretakers are not finely attuned to

the needs of a child, that child’s developmental processes are disturbed. The five

earliest stages of development are the most crucial to the adult personality that

eventually emerges. Those five stages are briefly reviewed below, because they are

central to the development and understanding of the PA:

1. The oral-sensory stage: trust vs. mistrust. In Erikson’s oral-sensory stage (so

named because the infant’s primary bodily zone for contacting the outside world is the

mouth), the psychosocial task is to establish a basic sense of trust in the outside world.

If the infant’s immediate world is gratifying and predictable, a sense of confidence is

created. The heart of this sense of trust is the quality of the mother-child interaction,

which the infant then extrapolates to the larger environment in which he/she operates.

Loving parents and a generally responsive environment teach the infant that he/she

lives in a world that can be trusted, a world in which help, care, and love are available,

a world that nurtures hope and a belief in the attainability of desires. In the case of a

failure in parental care, on the other hand, strong feelings of distrust are generally the

consequence.

9

2. The muscular-anal stage: autonomy vs. shame/doubt. A sense of mistrust leaves a

child ill-equipped to deal with the vicissitudes of the next stage—muscular-anal—

where the challenge is to acquire a sense of autonomy, to learn when to hold on and

when to let go. Effectively, it is the period of time when the child is learning to

control body functions (i.e., learning to use the toilet), but it also encompasses

growing self-sufficiency in many areas as the child gains mobility. If caretakers

encourage the child in that self-sufficiency, he/she develops a sense of will,

conscientiousness, and impulse control. Children who are not encouraged in this way

fail to acquire a sense of autonomy; without adequate support from their caretakers,

these children are burdened with an excessive sense of shame and doubt, they

question their worth, and they doubt their ability to exercise free will. Children who

feel that they have little control over their destiny often develop a sense of

helplessness, an attitude of laissez-faire and non-involvement.

3. The locomotor stage: initiative vs. guilt. The locomotor stage, with its focus on the

polarity of initiative versus guilt, is a time of still greater mobility and of the

exploration of time, space, and fantasy. Play is an important factor in practicing the

new freedom that increased mobility brings. Although much of a child’s play is

solitary, the ability to play is enhanced by the willingness of other family members to

engage the developing child—to enter his/her fantasy world (Winnicott, 1975). Such

encouragement contributes to the child’s possessing an adventurous attitude toward

life. During this period, children attempt to undertake many adult-like activities and

begin to learn how conscience governs initiative. When that exploration takes them

over the limits set by parents, they develop a sense of guilt. Anxious caretakers who

put a hold on the child’s natural playfulness exacerbate that guilt. Children held back

have a more prudent, conservative way of dealing with the environment at large,

while children allowed to engage in role experimentation develop a sense of purpose.

Caretakers also help to determine whether guilt is the dominant emotion as the child

experiences an awakening of sexual interest.

4. The latency stage: industry vs. inferiority. In almost every culture, children enter a

phase of development in which instruction is more or less institutionalized. The

challenge for young people is to develop a sense of industry rather than inferiority

during these school years. School offers a time of phenomenal learning and

10

competency acquisition, provided that feelings of inferiority are not internalized. If

the larger environment fails to recognize and affirm the child’s developing capacities,

he/she may experience what we might call “work paralysis.” The child who despairs

of his/her skills and status among peers may abandon assertiveness for a pattern of

self-effacement.

5. The adolescence stage: identity vs. role confusion. The previous stages of

development culminate at adolescence in the process of identity formation. A sense of

self-esteem is built on successive identifications, starting with the primary caretakers.

During adolescence young people must reconcile the changes in their body and social

position with these previous identifications. They must not only arrive at a sense of

who they are; they must also determine who they would like to be. This includes

questions about vocational role (“What career would best suit me?”) and sexual

identity (“What does it mean to be a man or a woman?” or “What kind of partner is

going to be most suitable for me?”). Some young people easily deal with the stress

that comes with this age, and their establishment of self is relatively unencumbered.

For others, this developmental task is overwhelming, leading to a sense of identity-

confusion, shaky self-esteem, and the urge to pursue questionable developmental

choices.

A sense of identity is at the core of a person’s being. It encompasses the experience of

his/her role in the world—his/her capabilities, values, and beliefs. Thus, although

biological factors may be at its core, identity is primarily a social construct. It relates

to the feeling that one’s life has a place among other lives; that one has value and is

taken seriously in the eyes of a meaningful group of people of which one is part; and

that one has the ability to manage the complex demands of daily life. Moreover, it

requires that each individual select and maintain goals that are personally meaningful.

A person’s identity is not written in stone, however. Although the basic building

blocks tend to be stable, identity continues to evolve during a person’s lifetime.

For the purpose of making sense of a person’s inner world, Erikson’s psychosocial

orientation toward development, as revealed in the above five stages, provides insight

into the forces that create motivational need systems (i.e., the forces that drive

individual behavior). Erikson’s success at creating a complex tapestry that weaves

11

together the various vibrant strands of human development enriches our

understanding of the “scripts” in the “inner theater” of individuals.

Motivational Need Systems

Erikson’s conceptual framework has been further developed by later contributions in

infant observation and research, and by neuro-physiological studies. Subsequent

researchers have recognized that mental processes evolve according to a biologically

determined timetable (Emde, 1981; Kagan, 1989; Kohlberg, 1981; Piaget, 1952;

White, 1959). Recognizing the delicate interface between nature and nurture, and

acknowledging a biological receptivity to external experiences at key points in the

infant’s development, they have emphasized the importance of the quality of early

experiences for adult psychic functioning.

Joseph Lichtenberg has proposed a number of motivational need systems that evolve

during the child–primary caretaker interface (Sullivan, 1953; Emde, 1981; Kagan,

1989; Lichtenberg, 1989; Lichtenberg, 1991; Lichtenberg, Lackmann, & Forshage,

1992; Lichtenberg & Schonbar, 1992;White, 1959). Each of these need systems can

be observed starting at infancy and continuing throughout the life cycle, altered by the

forces of age and by learning, maturation, and the responses of caretakers and other

loved ones. Each need system self-organizes or self-stabilizes, influenced by both

innate (hard-wired) and learned response patterns. Shifts in motivational-need

dominance that occur with age, development, and changing circumstances take

tangible form in emotional reactions. With each shift there is a developmental

resolution, based on whatever motivational needs are currently primary, and these

accumulated resolutions determine the make-up of one’s sense of identity as the self

becomes an independent center for initiating, integrating, and organizing experiences.

One motivational need system regulates a person’s physiological requirements, such

as the need for food, water, elimination, sleep, and breathing. Another system

encompasses an individual’s needs for sensual enjoyment and (later) sexual

excitement. Still another develops in response to the need to respond aversively to

certain situations through antagonism and withdrawal. There are, however, two

additional need systems that are of particular interest for life in organizations. The

12

first encompasses our need for attachment and affiliation; the second encompasses our

need for exploration and assertion.

The need for attachment and affiliation. Humankind’s essential humanness is found in

the seeking of relationships with other people, in feeling oneself to be a part of

something larger—that is, in what psychologists call attachment. Humans feel an

innate yearning for interpersonal relatedness, a universal craving to engage with

another human being, to be close to someone, to experience the pleasures of sharing

and affirmation. When this need for intimate engagement is extrapolated to groups,

we call the desire to enjoy intimacy a need for affiliation. The strength of this need

determines one’s position on the continuum of extroversion versus introversion (Jung,

1923). Both attachment and affiliation provide an emotional balancing role by

confirming the individual’s self-worth and contributing to his/her sense of self-

esteem.

The need for exploration and assertion. The other need system crucial to the work

world is the need for exploration and assertion. Associated with cognition and

learning, exploration involves the ability first to play and then to work. It develops

early in humans: researchers have found that novelty and the discovery of the effects

of certain actions cause prolonged states of attentive arousal in infants (Lichtenberg,

1989). The satisfaction that successful exploration brings continues into adulthood,

however. Closely tied to this need for exploration (building on Erikson’s observations

about the birth of initiative) is the need for self-assertion—that is, the need to be able

to choose what one does. As motivation, these related needs for exploration and

assertion produce a sense of effectiveness, competency, and self-efficacy (White,

1959, 1966; Bandura, 1986). In addition, playful exploration and manipulation of the

environment makes for a sense of autonomy, initiative, industry, and (again

extrapolating from Erikson’s work) adventurousness. Striving, competing, and

seeking mastery are fundamental motivational forces of the human personality, and

success in these arenas results in affirmation of the self.

Mood States. When we look at motivational need systems, we often see evidence of

another elemental aspect of personality: emotions. Nothing is more central to who a

person is than the way he/she regulates emotions. Along with cognition, emotions

13

determine behavior; and characteristic patterns of emotion, thought, and behavior

shape personality.

The emotional reactions of infancy are primarily biological, and they are tied to the

most basic of the need systems. From early on, however, socialization occurs through

the mediation of the primary caretakers. As socialization progresses, developmental

processes enable the individual to take on the various emotional “roles”—sadness,

joy, and so on. While all humans are born with a particular temperament, this

constitutional quality gives each individual only a predisposition to certain emotions.

Before a person is able to express any given emotion, the imagery associated with that

particular feeling state has to be internalized. Such internalization occurs as the child

grows and matures and learns from socialization. By the time adulthood is reached,

affect regulation has become an intricate part of one’s personality, and mood state can

be used as a barometer of psychological and physical well-being. The internalization

of imagery related to mood state should not be seen as static, however. Imagery can

be changed due to life experiences, making for different means of emotional

expression as a person matures.

The experiencing of emotions enables people to come into greater contact with

themselves, to find out what they feel (as opposed to think) about things, what they

like and dislike, and what they want and don’t want. Some people are able to express

emotions appropriately and comfortably, while others struggle to find words for what

they feel, and associate emotions (sometimes even those that we think of as positive)

with painful thoughts. Emotions color experiences with positive and negative

connotations, creating preferences. They also serve people in many adaptive and

defensive ways, depending on one’s personal “script.” Although there is a wide range

of emotions, only a few mood states seem to account for a large number of feelings.

Most mood variations can be explained by just two factors, which can be labeled

simply positive affect and negative affect. As a result, a number of scholars have

mapped emotions into an affective space determined only by affect evaluation

(negative-positive) and affect activation (aroused-unaroused) (Darwin, 1920; Ekman

& Davidson, 1994; Lazarus, 1991; Plutchik, 1962; Tomkins, 1995; Watson, Clark, &

Tellegen, 1988). This simple method of mapping emotion can be extremely effective.

14

The Personality Audit, as an assessment of psychodynamic functioning, would be

incomplete without reference to the role of emotions in the makeup of the person.

Consequently, expressiveness of mood state is included as one of the dimensions,

indicating the intensity of positive and negative arousal. Mirroring the simplicity of

the emotion-mapping described above, the scale measures the extent to which the

person feels high-spirited or low-spirited.

The various theoretical underpinnings of today’s understanding of human

development that have been discussed above form the conceptual basis for the

constructs of the PA. The seven dimensions of that instrument—derived from basic

aspects of personality—can help us understand the vicissitudes of personality

functioning. These dimensions, when assessed by self and others, give us a quick

glimpse into the test-taker’s inner world.

The Advantages and Disadvantages of Dimensional Models

As noted earlier, the PA is based on a dimensional model of personality traits

(meaning that people are scored according to where they lie on a continuum on each

of various different dimensions). Dimensional models have an advantage over so-

called categorical models in that they present personality fluidly, without discrete

boundaries of what is considered normal and abnormal. Furthermore, they allow for a

rich representation of individuality rather than forcing people into specific categories;

by permitting various combinations of a number of clinical features or personality

dimensions into a single profile, they encourage breadth and comprehensiveness.

Because they give no single dimension pride of place, they “lose” less information

than models dependent on discrete traits.

Dimensional models are not without methodological problems, however. Given the

nature of human development, the dimensions are rarely independent; rather, they

build upon each other, making psychometric assessment more difficult. Furthermore,

any organizing system—however extensive the attempt at full coverage of the various

aspects of personality functioning, and whatever the number of dimensions—implies

a restriction in the presentation of the richness of personality. It represents, at best, a

15

window (or one view) into the very complex inner theater of the individual, and

psychologists disagree on how many dimensions are needed to create a large enough

window to see accurately (Cattell, 1957; Eysenck, 1960; Hogan, Johnson, & Briggs,

1997).

Perhaps the most troubling limitation inherent in the dimensional approach is that,

while the dimensional scale gives no single dimension pride of place, as noted earlier,

test-takers and test-scorers inevitably interpret some personality traits as being more

positive than others. Some of the “anchor points” that delineate the dimensions are

typically perceived as being more or less desirable than others. To minimize this

problem in the PA, we have made an effort to present the anchor points as neutrally as

possible.

But even being positioned on what is often perceived as a desirable point on the

dimension can have its downside. The dimensional approach suggests that it is

possible to have too much of a good thing. In other words, although one side of a

specific dimension may appear more attractive, an extreme position on that dimension

can be problematic. For example, while a rating in the direction of negative self-

esteem is clearly troubling, a rating at the extreme position of positive self-esteem is

not without problems of its own. It may indicate that the test-taker is not simply

confident, but is convinced of his/her superiority. Likewise, while a rating toward

low-spiritedness indicates potential depressive problems, an extremely high rating can

suggest a manic state. These curvilinear concerns apply to all the dimensions, making

it hard to label as “right” or “wrong” any position on the dimensions. This concern

embodies the limitations—whatever approach is taken—of using a test to assess

personality, in all its richness.

Another reason that we chose a dimensional approach for the PA was our belief that

any such instrument must be ecologically valid (Messick, 1994)—that is, its findings

must be generalizable and transferable to the environment in which the tested

behaviors flourish (in this case, the business environment). In other words, the

interpretation of the results, and the communication of this interpretation to the person

tested, is necessary to the validation process. For this reason, we chose many

correlated dimensions rather than fewer orthogonal ones, believing that their ease of

16

understanding by “laypeople” and their relevance for personal development in a

leadership context would render them particularly ecologically valid. This is

especially important in regard to the PA, because it is intended for use in training and

executive coaching contexts, where it is crucial that the people involved can

understand and apply the results.

Users of the PA need to keep these various constraints in mind, remembering that the

goal of assessment using this instrument is arriving not at a simple set of scores, but

rather at a better understanding of one’s personality—in all its complexity.

The Seven Personality Dimensions The PA is based on various conceptualizations of personality development, as the

earlier discussion suggests. Our first step, in creating the instrument, was to identify

those personality dimensions that broadly organize the domain of human functioning

and that are most helpful in explaining executive behavior. We also looked at specific

examples of various leaders for evidence of their salient personality traits (Kets de

Vries, 1993, 1995, Kets de Vries & Florent-Treacy, 1999, Kets de Vries, Shekhnia,

Korotov, & Florent-Treacy, 2004). Then we drafted a comprehensive description of

each of the dimensions identified and wrote a large pool of assessment “items” (the

statements to which test-takers respond), using the descriptions as guidelines. After

testing the items for face validity (that is, making sure they made sense to test-takers),

we weeded out about half the items, leaving us with six items per dimension.

Two anchor points—for example, negative self-esteem and positive self-esteem—

uphold each of the seven dimensions of personality assessed in the PA. A set of

specific questions assesses each dimension, with the test-taker responding somewhere

on the continuum between the two anchor points. The dimensions are listed opposite

one another on the assessment, followed by statements that characterize people who

favor each anchor point.

Let’s look here at each of the seven dimensions in turn, to see what personality traits

are assessed by each:

17

Negative Self-Esteem – Positive Self-Esteem. The first dimension we consider—the

result of a complex developmental process—centers on identity formation. Identity is

the outcome of each individual’s highly subjective struggle to develop a sense of inner

sameness and continuity and to articulate the role of the self vis-à-vis the external

world (Erikson, 1963, 1968; Westen & Heim, 2003). The label self-esteem reflects

how the person measures up to a desired self-image. Self-esteem reflects an evaluative

self-judgment based on self-knowledge (Baumeister, 1998). Although researchers

argue about whether to treat self-esteem as a relatively stable trait (a person’s general

or most typical feelings about him/herself) or as a temporary state (a person’s feelings

at a certain moment in time), there appears to be a strong correlation between trait and

state self-esteem, the implication being that most people tend to have a stable self-

esteem baseline (Kernis, 1993; Heatherton & Polivy, 1991; Baumeister, 1998; Leary

& MacDonald, 2003). To the extent that what one sees in oneself approximates what

one would like to be, self-esteem is positive, while negative self-esteem reflects

dissatisfaction with one’s desired self-image. People who score high on positive self-

esteem radiate self-assurance and self-confidence (although inflated self-esteem may

be an expression of excessive narcissism), while people who score toward negative

self-esteem come across as insecure. Self-esteem may serve as a foundation for

development of one’s assertiveness in life, as well as one’s mood state.

Examples of items measuring the Negative Self-Esteem – Positive Self-Esteem

dimension include:

I think other people find me…

boring – extremely interesting.

Looking at myself …

I am self-critical – I accept myself fully.

Low-Spirited – High-Spirited. No study of personality would be complete without a

focus on mood state. Mood refers to pervasive and sustained emotion that is

subjectively experienced. It can be pleasurable (such as elation) or unpleasant (such as

depression or anger). Mood state colors a person’s perception of the world. It also

serves as an internal and external signaling system, indicating to others the emotional

18

state that the person is in. People who fall toward the high-spirited end of the

continuum are characterized by intensity of reaction. They display strong emotions

and a high degree of expressiveness, and they may be subject to violent mood swings.

People who tend to be low-spirited, on the other hand, are characterized by flat,

shallow, constricted, changeable, or irritable affective expression. As mentioned

above, the mood state may be based on one’s self-esteem.

Examples of items measuring the Low-Spirited – High-Spirited dimension include:

I am optimistic…

rarely – almost always.

I feel hopeless…

often – rarely.

Distrustful – Trustful. The Distrustful – Trustful dimension, described by Erikson

(1963, 1968, 1975) as the first “psychosocial” personality dimension (and the

developmental goal of the first of his life stages, discussed earlier), is the fundament

on which all the other psychosocial characteristics (Erikson’s stages 2 through 5) are

built. In seeking the roots of this dimension, Freud (1905) takes us back to the oral

phase, that stage in life at which we form the basis for trusting and affectionate

relationships with others. Basic trust of others makes for a sense of mutuality and

creates an optimistic outlook toward life and fellow humanity. People high on trust

tend to be considerate of others and possess a hopeful attitude toward life experiences

(although excessive trust results in a potentially dangerous naiveté). Individuals high

on trust have also been described as being more inclined to offer others a second

chance and to interweave their economic exchange with social exchange (Husted,

1989; Ring, 1998). The social-exchange component in the behavior of individuals

high on trust suggests that this dimension is related to one’s openness to social

interactions in general, as expressed, for instance, by a tendency to be more

extroverted. People high on distrust, on the other hand, are watchful about perceived

dangers in the environment, and they seem more distant, more guarded, and more

cold-hearted. In economic-exchange situations, people high on distrust have been

described as requiring a great deal of specificity in laying out the terms of the

19

exchange (Husted, 1989; Ring, 1998). Interestingly, the tendency to trust or distrust

others may correlate with another important personality dimension—the degree to

which one is prudent versus adventurous. Adventurousness presupposes a certain

degree of trust toward life situations and the actors involved in them; therefore, we

might expect that people high on trust would be, generally speaking, more

adventurous.

Examples of items measuring the Distrustful – Trustful dimension include:

When people hurt me…

it is difficult for me to forgive them – it is very easy for me to forgive them.

I reveal myself to others…

very little – completely.

Introverted – Extroverted. This dimension is made up of the polarity of introversion

versus extroversion, reflecting the way that people relate to the external world (Jung,

1923). People who fall toward the extroversion end of the continuum tend to direct

their energy toward the external world, orienting themselves toward people and

external situations. Gregarious, uninhibited, impulsive, attention-seeking, and

possessed of great interpersonal charm, they tend to be very effective in social

settings. People toward the introversion end of the continuum, on the other hand,

orient themselves more toward their inner world. Their energy flows inward first, then

toward the outer reality. Because they present a much lower profile, they tend to be

viewed as rather shy, reserved, and cautious. At times, their behavior may be

interpreted as unsociable. These people may be viewed as not trusting others and as

cautious and non-adventurous, because they are not oriented toward engagement with

others. The Introverted – Extroverted dimension is often included in well-known

personality measures, such as the Big Five model of personality (e.g., Costa &

McCrae, 1980; McNulty, 2000), where this dimension is reflected in the Extraversion

or Surgency element.

Examples of items measuring the Introverted – Extroverted dimension include:

I would prefer to spend most of my time…

20

alone – with other people.

I seek the company of other people…

rarely – quite often.

Prudent – Adventurous. This dimension refers to how adventurous a person is. It is

closely tied to the exploratory motivational need system, discussed earlier

(Lichtenberg, 1991)—a need system that is activated in infancy, with a baby’s

capacity for learning and for cognitive and intellectual development. People who

score high on adventurousness tend to be unconventional, imaginative, creative,

inventive, artistic, and eager to experiment with new things. Individuals who score

toward the prudent end of the continuum tend to be more conservative, conventional,

and conforming. Lacking curiosity and imagination, they are content with the status

quo and prefer to avoid risk-taking. While adventurousness is often based on a sense

of inner security, it sometimes reflects the overly rebellious streak of someone out to

prove that he/she can make a difference. Thus having a more prudent orientation can

be, in some instances, a sign of greater mental health. This dimension is relatively

closely related to the Distrustful – Trustful dimension and to the Introverted –

Extroverted dimension. Furthermore, the Prudent – Adventurous dimension may be

related to the Openness to Experience or Intellect factor of the Big Five model, which

deals with such characteristics as imagination, curiosity, artistic expression, insight,

and sophistication. However, it has been argued that in the Big Five model (McNulty,

2000) this dimension is not clearly defined. The Prudent – Adventurous dimension,

which covers a part of the Openness to Experience factor in the Big Five model,

seems to be a more suitable construct, because it deals with a narrower set of personal

characteristics.

Examples of items measuring the Prudent – Adventurous dimension include:

In my life I need a great deal of…

stability – variety.

I seek new thrills…

rarely – very often.

21

Laissez-faire – Conscientious. The Laissez-faire – Conscientious dimension draws

on the second of Erikson’s (1963, 1968, 1975) stages of development—the muscular-

anal stage—and on what Freud referred to as the anal phase (Freud, 1905). This

dimension addresses the abilities of self-motivation and self-control. People who

score high on conscientiousness like structure, prefer to take personal responsibility,

generally behave in an orderly manner, and are systematic, methodical, and efficient.

They are models of self-control. This dimension is close to the Conscientiousness

dimension in the Big Five model of personality. Researchers have shown that

conscientious individuals often seek and achieve social approval and acceptance-

gaining outcomes (Leary & MacDonald, 2003). What they lack in spontaneity, they

make up in thoroughness and attention to detail. On the other hand, people with a

laissez-faire personality have a happy-go-lucky attitude. Typically unconcerned about

rules, regulations, and details, they tend to be rather flexible; however, they also tend

to let things slip. For these reasons, laissez-faire individuals may be valued less as

relational partners and group members (Leary & MacDonald, 2003).

Examples of items measuring the Laissez-faire – Conscientious dimension include:

When I don’t do what I promised,…

I don’t worry about it – I feel guilty.

I pay…

little attention to details – great attention to details.

Self-Effacing – Assertive. In discussing traits assessed on the Self-Effacing –

Assertive dimension, Freud (1905) spoke of the Oedipal phase, while Erikson (1963,

1968, 1975) presented the polarity of initiative versus guilt. When children develop

greater mobility and engage in exploration of time, space, and fantasy, they explore

the new freedom brought about by their increased mobility. Children attempt to

undertake adult-like activities and start to learn how conscience governs initiative.

With limits set by parents, children begin to experience the feeling of guilt when they

exceed those limits. The Self-Effacing – Assertive dimension encompasses our

competitive strivings. Assertiveness is about satisfaction of the need to choose what

one does. Although some people have argued that assertiveness is not a generalized

personality dimension (Fischer, 1987), it is still primarily viewed as a trait or a set of

22

traits or elements of personality (Twenge, 2001). People high on assertiveness deal

with the world with a sense of purpose. Such individuals have high aspirations and are

both action-oriented and achievement-oriented. Hardworking and tough-minded, they

like to test their abilities by competing with others, and they like to win. Both

laypeople and professional psychologists tend to agree that assertiveness includes

standing up for one’s rights, freely expressing opinions and feelings, being sure of

oneself, and being a leader (Twenge, 2001). People who are closer to the self-effacing

end of the continuum tend to be reflective, weighing all options before coming to a

decision. They tend to be less ambitious and more socially reticent than their assertive

peers.

This dimension is related to the self-esteem dimension of personality, because people

who are high on self-esteem may be expected to be more assertive, while those low on

self-esteem may be expected to be low on assertiveness.

Examples of items measuring the Self-Effacing – Assertive dimension include:

I defend my point of view…

rarely – almost always.

For me, winning is…

unimportant – extremely important.

Validation Study

The reliability of the PA was studied in two stages: first, by checking the internal

consistency within each scale using a classic psychometric approach; and second, by

evaluating the structure of the questionnaire through factor analysis at both the item

level and the scale level.

The questionnaire includes 42 items spread over seven scales with six items each.

Each item is presented in a bipolar form proposing two opposites statements.

Instrument-takers are asked simply to indicate on a seven-point Likert-type scale the

degree to which the left or the right pole of each statement describes the way they (or

23

the individuals that they, acting as observers, are assessing) act in a particular

situation. As a guideline, they are advised that the scale has a 4 in the middle and goes

up from 1 – 3 to the left and 5 – 7 to the right. On each side of the scale, there is a

statement that describes self-perceptions. The two descriptions are opposites. The

subjects have to read each statement, decide how they feel about it, and then mark the

number that describes them (or the individuals they are observing) best. If they feel

that their behavior is somewhere in the middle of the two contrasting statements, they

should mark 4. After the questionnaire has been completed by the test-takers and all

invited observers, and the results have been compiled, the test-takers receive a

printout as shown in Table 1.

Table 1: Presentation of printout of The Personality Audit

Negative Self-Esteem

Distrust

Laissez-faire

Self-Effacement

Introversion

Low Spirited

Traditional

Positive Self-Esteem

Trust

Conscientiousness

Assertiveness

Extraversion

High Spirited

Adventurous

-20 -15 -10 -5 0 5 10 15 20

The Personality Audit

Sample

The sample used for the validation study of this survey instrument was constructed

using data from 23 groups of executives from a wide range of nationalities and

24

cultures who attended executive education programs at INSEAD in 2002, 2003, and

2004 and completed the questionnaire. While in early samples individuals used a

paper-and-pencil version of the instrument, the PA was later turned into an Internet-

based assessment tool. Participants could access the “Self” questionnaire and specify

observers from whom they were seeking feedback (along with the observers’ e-mail

addresses). The system would then send an e-mail message to observers and inform

them that their input was requested. A link to the website with the questionnaire was

provided.

The questionnaire was prepared in two versions: “Self” and “Observer.” The

Personality Audit “Self” version was completed by 617 subjects. In addition, 549 of

them also asked an average of three (and a maximum of ten) observers to complete

the PA as their observers, thus allowing us to test the 360-degree feedback component

of the instrument. The observers were classified within the following categories: peers

(130), subordinates (314), superiors (206), partners (276), family members (81),

friends (200), secretaries (25), not specified (682). The sample used for this validity

study comprises, therefore, 2531 questionnaires (617 “Self” audits and 1914

“Observer” audits).

The division between genders was unbalanced in favor of males for the “Self” version

of the instrument (68%) and quite well balanced for the “Observer” version of the

instrument (54% males, 46% females). The mean respondent age was 42 years for the

“Self” version (std = 8 years, age minimum = 21 years, maximum = 68) and 42 years

and three months for the “Observer” version (std = nine years and three months, age

minimum = 10 years1, maximum = 74). The respondents belonged to 57 different

nationalities, representing European and North American countries primarily (British,

30%; German, 10%; French, 7%; US, 6%; Swedish, 6%; Finnish, 5.5%; Dutch, 5%;

Russian, 5%; Belgian, 3%; Danish, 3%). These relative frequencies are roughly

equivalent for both “Self “ and “Observer” versions.

Results 1 This very young age and six others below twenty, are indicated for “family” observers; we can make the hypothesis that some people asked their own children to answer the questionnaire.

25

Internal Consistency Analysis

The means and standard deviations appear in columns 4 and 5 of Table 2. The

theoretical maximum score per scale is 42. The average scores are higher than the

theoretical mean (21) in general, indicating a ceiling effect due partly to the social

desirability factor. This finding, once again, confirms the need for a multiple feedback

component in order to provide a more realistic picture for executives taking the

assessment. These high-score issues will be explored and discussed later.

Table 2: Means, standard deviations, and Cronbach’s alpha for the scales of the PA estimated on the whole sample (N = 2531)

Scale

Low-score pole

Scale

High-score pole

Abbreviation Mean Standard

Deviation

Cronbach’s

Alpha

Negative Self-

Esteem

Positive Self-

Esteem

NSE_PSE

28.97 4.12 0.70

Distrustful Trustful DIS_TRUS 26.75 5.41 0.70

Laissez-Faire Conscientious LF_CONSC 30.66 5.01 0.69

Self-Effacing Assertive SE_ASSER 30.54 4.72 0.72

Introverted Extroverted INT_EXTR 27.94 5.60 0.78

Low-Spirited High-Spirited LS_HS 32.10 5.20 0.77

Prudent Adventurous PR_ADV 29.05 5.55 0.79

Internal reliabilities, assessed through standardized Cronbach’s alpha (see column 6,

Table 2), range from .69 to .79 for the different scales of the PA. The lowest values

are close to the .70 value generally considered to indicate a sufficient reliability by

classical psychometric treatises (for example, Nunnally, 1978) and by standard

practice within the scientific community (Peterson, 1994). These low values are

partially due to the small number of items per scale (6). In fact, research has found

that the number of items in the calculation of alpha coefficients can appear to create

confusion between internal consistency and the length of the scale (Cortina, 1993).

This small number of items presents, however, the advantage of a short time for the

administration, an important characteristic for a 360° instrument that must be

completed by busy respondents. The downside is that the values for reliability as

estimated through Cronbach’s alpha are not as high as one would like them to be,

26

although most of the items present sound psychometric properties. Furthermore, the

Corrected Item/Total Correlations (CITC) has been computed for each item. The

CITC mean value (.47) indicates that most of the items present a high relationship

with their scale. Only one item with a low CITC (.08) seems problematic.

The internal reliability of 360-degree feedback instruments is, in general, lower for

questionnaires filled out by the subjects themselves than for questionnaires filled out

by their “observers” (Posner & Kouzes, 1988, 1993). It is hypothesized that the

difference in internal consistency can be explained by the manifestation of the social

desirability factor. Another hypothesis can be an artificial inflation of the reliability

indicators due to the data collection design. Several observers rate the same subject in

the “Self “position, and the resulting data are structured as a nested design (with the

raters nested in the observers). This implies that there exist some dependencies among

multiple observers rating a single self. This situation creates a violation of the

assumption of local independency under which the reliability indexes are valid (Lord

& Novick, 1968).

To avoid this problem we tested a sample containing only one observer per assessed

test-taker by drawing one observer at random for each “Self“ version that had several

corresponding “Observer” versions. Cronbach’s alphas were then computed

separately for the “Self” and “Observer” questionnaires. Through that calculation, we

found that reliability ranges from .62 to .78 for the “Self” scores (column 2, Table 3),

and from .68 to .80 for the “Observer” scores (column 3, Table 3). Thus the

observations of observers appear to be more reliable than the self-ratings. This

phenomenon, undoubtedly related to a more reliable use of the response scale by

observers, underscores the importance of using multiple feedback instruments to help

executives gain a better understanding of their behavior.

We also assessed self/observer reliability via an inter-rater reliability approach,

finding the intraclass correlation coefficient (Shrout and Fleiss, 1979). Because the

number of observers was not the same for every subject (varying from 0 to 7), the

sample made up of only one observer, selected at random, for each self (as described

above) was used. The reliability of the inter-rater agreement can be foreseen from two

generalizability situations: the use of the rater’s single score (comparison between self

27

and each observer) and the use of the average of several raters’ scores. The formulas

used to estimate the inter-rater reliability were different for both situations. Because

the PA can be used for both situations, the ICC values have been computed for the

single situation (Table 3, column 3) and the average situation (Table 3, column 4).

These values indicate a reliable agreement between self and observer ratings. The

lowest values are observed for Low Self-Esteem – High Self-Esteem and Prudent –

Adventurous dimensions.

Table 3: Reliability indexes based on PA “Self” and a random sample of

“Observers”

Cronbach’s Alpha

Self† Observer$ ICC$

Single

ICC$

Average

NSE_PSE 0.71 0.71 0.18 0.57

DIS_TRUS 0.62 0.72 0.27 0.69

LF_CONSC 0.67 0.68 0.19 0.59

SE_ASSER 0.72 0.69 0.14 0.50

INT_EXTR 0.75 0.78 0.28 0.70

LS_HS 0.75 0.80 0.33 0.75

PR_ADV 0.78 0.80 0.36 0.77

† n=617

‡ n=549 $ Internal Consistency Coefficient, case 1 allowing to Shrout & Fleiss, 1979

Cronbach’s alpha gives information on the reliability of the scale as a whole. It is

equally important, however, to look at internal consistency at the item level. The

strength of the relationship of each item to its scale can be measured by examining the

Corrected Item Test Correlation (CITC). In the PA, the CITC values range, for the

whole sample, from .08 to .67, with a median at .49. For 33 items out of the 42, the

CITC is higher than .40, a value usually considered to indicate a reliable relationship

between the item and the scale to which it belongs; for 7 items the values range

28

between .30 and .40. Only two items show a CITC below .20. (These items will be

improved in further versions.) Based on this analysis, we can conclude that 95% of

items are well placed in their respective scales.

Factor analysis

Factor analysis at the item level. In order to test the hypothesis that a seven-

dimensional model can explain the relationship among the items, the structure of the

questionnaire was studied with exploratory factor analysis (EFA), using the

procedures implemented in LISREL 8.5 (Jöreskog & Sörbom, 2000). This procedure

extracts the number of factors fixed (seven in this analysis) using a principal factor

analysis of the polychoric correlations matrix (since the items are seven-point Likert

scales); the factor extraction is then followed by promax rotations. Separate analyses

were performed on the “Self” (n=617) and “Observer” (549, drawn at random)

samples. Both analyses brought very similar results concerning the fit of a seven-

factor model following the root mean square error of approximation (RMSEA) index

(Browne and Cudeck, 1993): RMSEA = .054 and RMSEA = .058 respectively for

“Self” and “Observer” samples. An examination of the highest loadings (e.g., >.40) of

the items for each factor shows the grouping on the same factor of the items

pertaining to the same scale. These groupings confirm that the theoretical structure

appears to reflect accurately the organization of the items for both samples. Few items

present a low loading on the factor on which the items pertaining to their scale are

grouped and a high loading on another factor. However, four items do not properly fit

the theoretical structure:

• Item 4 theoretically belonging to the Self-Effacing – Assertive scale presents a

loading on the factor incorporating the Negative Self-Esteem – Positive Self-Esteem

scale items (in the “Observer” sample only);

• Item 24 theoretically belonging to the Low-Spirited – High-Spirited scale

presents a loading on the factor incorporating the Distrustful – Trustful scale items (in

the “Observer” sample only);

• Item 25 of the Self-Effacing – Assertive scale presents a loading on the factor

incorporating the Distrustful – Trustful items (on the “Self” scale only);

29

• Item 29 of the Distrustful – Trustful scale presents a high loading on the factor

incorporating the Low-Spirited – High-Spirited items.

These items have been previously identified as less reliable in the above-mentioned

homogeneity analysis. The multiple belonging of some items may represent an artifact

inflating the relationships between the scales.

Concerning the relationship between the scales, it is important to remember that the

Promax procedure used in our analysis is a rotation method that allows factors to be

correlated. The inspection of the correlation matrix between factors shows significant

links among the dimensions measured. These correlations are particularly high among

certain groups of dimensions, which indicates the regrouping of certain scales at a

higher level. This point will be developed further when we discuss factor analysis at

the scale level.

Factor analysis at the scale level. A combination of exploratory and confirmatory

factor analysis (CFA) using a specification search approach was used for two main

reasons. First, the structure of relationships among the seven scales could be only

partially determined with the help of theories. Second, testing the fit of only one

model to the data is often not very heuristic, and is sometimes unrealistic. The

principal argument advanced against this strategy of testing one unique model is the

existence of equivalent models (models presenting the same number of parameters

and the same fit indexes; on this point see, in particular, MacCallum, Wegener,

Uchino, & Fabrigar, 1993). Another approach, specification search, which relies on

the test of several apparently plausible models, seems more heuristic. In advocating

use of a set of plausible models, McCallum (1986) distinguishes between two kinds of

relationships between observed and latent variables: the obligatory and the optional.

Obligatory relationships imply that the model cannot be conceptualized without

containing these elements. These obligatory relationships need to be present in all

models that are being tested. Optional relationships, on the other hand, can improve

the model, even in the absence of a definitive hypothesis about their simultaneous

presence in the model. The test of the model relies on a theoretical approach based on

the concepts of the Kullback-Leibler (K-L) information theory. (For a recent review

and applications to model search see Burnham and Anderson, 1998, and for its

30

application to structural equation modeling see Raftery, 1993). The K-L best model

gives the maximum of information considering to the data that is being used while

creating the best fit for whatever parameters are used in the model. This approach,

used in structural equation modeling, has been implemented in the analysis of

covariance MOmentS (AMOS) model, version 5.0 (Arbuckle & Wothke, 1999;

Arbuckle, 2003).

Because it was necessary to avoid the dependencies existing between a test-taker and

his/her observers, four independent samples were determined through a random

sampling: two “Self” samples and two “Observer” samples. To construct the model

used for specification search, the number of factors was determined through EFA

conducted on the first “Self” sample. This analysis led to the retention of four factors.

From a theoretical viewpoint, a four-factor solution seemed a good option, because

three of the scales are similar to elements of the Big Five personality model (e.g.,

Costa & McCrae, 1980): Introverted – Extroverted (Extraversion of the Big Five

model), Laissez-Faire – Conscientious (Conscientiousness of the Big Five model) and

Self-Effacing – Assertive (Agreeableness of the Big Five model). A fourth, the Low-

Spirited – High-Spirited dimension, is made up of items that can be considered

relatively close to the Neuroticism factor of the Big Five personality model. These

four scales can be considered as four independent factors. For the specification search

procedure, the relationship between each of these four scales and one factor was

deemed obligatory. The relationships among these four factors, however, and the

three remaining scales (Negative Self-Esteem – Positive Self-Esteem; Distrustful –

Trustful; Prudent – Adventurous) have been modeled as optional. AMOS 5.0 was

used to operate the specification search on the first “Observer” sample. Among these

models, one was retained for its well-fitting properties and its relatively simple

structure (see loadings in Table 4). In a second stage, this model was fitted on the two

samples (“Self” and “Observers”) that had not been previously used in the model

construction. The results demonstrate an acceptable fit for the “Self” and “Observer”

samples. The values of the loadings are quite similar to those observed on the

construction sample.

Table 4: Loadings of the seven scales on the four factors for the validation sample

31

Factor 1 Factor 2 Factor 3 Factor 4

NSE_PSE .36 .41

DIS_TRUS .84 -.50

LF_CONSC 1.00

SE_ASSER .87

INT_EXTR .60

LS_HS 1.00

PR_ADV .68

The loadings values, presented in Table 4, lead to identification of four domains of

behaviors assessed through the PA. The first factor groups together the Low-Spirited –

High-Spirited scale and the Negative Self-Esteem – Positive Self-Esteem scale. This

factor can be interpreted as dealing with the person’s mood, either positive or

negative, and with an evaluation of the self as compared to the desired self. The

second factor underlies the Introverted – Extroverted, Distrustful – Trustful, and

Prudent – Adventurous scales. This factor can be interpreted as dealing with behaviors

related to extroversion, sociability, and change. The third factor groups together three

scales: Self-Effacing – Assertive, Negative Self-Esteem – Positive Self-Esteem, and

Distrustful – Trustful (negative loading). This factor deals with dominant social

behavior or assertiveness. A negative relationship with the Distrustful – Trustful scale

indicates that it is bound to the distrustful pole of the items. The Positive Self-Esteem

linkage probably implies, in the case of high scores, a superiority in the social

comparison (“I’m better than other people”). The fourth factor loads only the Laissez-

Faire – Conscientious scale.

These factors are oblique, but only three correlations have been integrated in the

model:

1) The correlation between F1 and F2 (.61) indicates that both these factors are

redundant for about a third of their variance. This might be explained by the fact that

they assess the positive mood and the warmth relationship with other.

32

2) The correlation between F2 and F3 (.42) might reflect the fact that both factors are

socially oriented.

3) The correlation between F3 and F4 (.23) indicates a low relationship between two

factors that relate to scales (and therefore items) dealing with people who believe in

affirming themselves and demanding a lot for themselves.

This analysis brings arguments to validate the theoretical structure of the PA as

constituted by four main underlying factors: mood, extroversion and sociability,

assertiveness, and conscientiousness. The main problem seems to be the medium

loadings of LS on the first and second factors, which convey the idea of a dichotomy

in the set of items: one being more related to positive mood, the other to social

warmth and sociability. This point should be underscored for the professional using

the PA.

Discussion

Analysis of the structure of The Personality Audit has identified four major factors

influencing the behavior of executives in both public and private life. The four

factors—mood, extroversion and sociability, assertiveness, and conscientiousness—

comprise seven dimensions of an individual personality: Low-Spirited – High-

Spirited, Negative Self-Esteem – Positive Self-Esteem, Distrustful - Trustful,

Introverted - Extroverted, Prudent – Adventurous, Self-Effacing – Assertive, and

Laissez-Faire – Conscientious. Data from this study suggests that these seven

personality dimensions possess sufficient internal reliability and consistency.

The fact that there are significant links among the various personality dimensions

measured may be viewed as a support of the theoretical underpinnings of the PA:

many of the behaviors demonstrated by test-takers are in fact coming from the same

biosocial, cognitive, and psychosocial foundations. This is also indirectly supported

by the similarity among some of the factors and elements of the Big Five personality

model. The difference, however, is that the dimensions measured by this instrument,

without doubt capturing only a fraction of the richness of a human personality, reflect

the areas that seem to be particularly important for the functioning of executives in

33

organizations. The objective of the PA is to deepen the test-takers’ awareness of what

makes them tick and of how others perceive what drives them. This instrument not

only helps executives start a journey of self-exploration, but also encourages them to

undertake a meaningful discussion with the people around them about the way those

others perceive them. The PA is also an effective tool for individual or group

executive coaching. (For a discussion of use of the PA in executive coaching sessions

see Kets de Vries (2004, 2005a, 2005b).)

Obviously, there is a need for further analysis of the psychological significance of

correlations among the dimensions loading on the same factors using factor analytic

and other multivariate techniques. This task, however, is beyond the scope of the

present working paper and remains an area for future research. Looking deeper into

each of the dimensions and the nature of their relationships with other personality

manifestations is a fascinating field, and one important for a better understanding of

the human mind and executive functioning.

Another exciting area for further research concerns addressing the differences

between the “Self” and “Observer” scores. We have indicated the difference that is in

general typical for 360-degree instruments. However, further research in differences

in perception of various personality dimensions may shed light on discrepancies

between public and private selves, as well as personality traits that are more or less

accessible to an outsider. The implications would include such important issues as

selection and development based on the observations of an individual.

Further research in differences along various dimensions measured by the PA is

important for increasing our knowledge about the influence of nationality, gender, and

age on the scores obtained. In particular, it would be interesting to see how

perceptions differ among various national cultures and between genders.

Another important area for future research involves comparing the results of the PA

with some established measures of managerial or leadership behavior, such as the

Global Executive Leadership Inventory (Kets de Vries, Vrignaud, and Florent-Treacy,

2004; Kets de Vries, 2004b, 2004c). Understanding the correlations between The

34

Personality Audit and the Global Executive Leadership Inventory could increase our

understanding of the driving forces behind certain leadership practices.

Executives must heed how others perceive them, because organizational decision-

making has to be reality-based if it is to be effective. But as myriad examples of

executive derailment have shown, people have a hard time dealing with reality. As T.

S. Eliot once said, “Humankind cannot stand very much reality.” Reality is not an

enemy, however: it is not reality that hurts, but rather our perception of reality. In life,

too many issues are decided on the basis of hate, love, lust, rage, envy, sorrow, joy,

hope, fear, and illusion. Individuals and organizations that make astute use of The

Personality Audit can keep such emotions from obscuring reality.

35

References Arbuckle, J.L. (2003). AMOS 5.0 Update to the Amos user's guide . Chicago IL : SmallWaters Corporation Arbuckle, J.L., & Wothke, W. (1999). Amos 4.0 user's guide . Chicago : SmallWaters Corporation Balint, M., Ornstein, P. H., & Balint, E. (1972). Focal Psychotherapy. London, UK: Tavistock. Bandura, A. (1986). Social Foundations of Thought and Action: A Social Cognitive Theory. Englewood Cliffs, NJ: Prentice Hall. Baumeister, R. (1998). The Self. In D. Gilbert, Fiske, S., and Lindzey, G. (Ed.), The Handbook of Social Psychology (4th ed., Vol. 1, pp. 680-740). Boston: McGraw-Hill. Bland, C., Edwards, M., & Kuhi, L. (1994). Academic administrator evaluation through colleague feedback. CUPA Journal, Spring,19-29. Bowlby, J. (1969). Attachment and Loss (Vol. 1: Attachment). New York, NY: Basic Books. Browne, H.W., and Cudeck, R. (1993). Alternative Ways of Assessing Model Fit. In: Bollen, K.A., and Long, J.S. (eds) Testing Structural Equation Models. Newbury Park, CA: Sage, 136-162. Burnham, K.P., & Anderson, D.R. (1998). Model selection and inference. A practical information-theoretic approach . New-York : Springer Verlag Carlson, M.S. (1998). 360-degree feedback : the power of multiple perspectives. Popular Government, 63(2), 38-49. Cattell, R. (1957). Personality and Motivation Structure and Measurement. New York, NY: New World Book. Cattell, R. B., Eber, H.W., & Tatsuoka, M. M. (1970). Handbook for the Sixteen Personality Factor Questionnaire. Champaign, IL: Institute for Personality and Ability Testing. Church, A., and Bracken, D. (1997). Advancing the State of the Art of 360-Degree Feedback: Guest Editors’ Comments on the Research and Practice of Multirater Assessment Methods, Group and Organization Management, 22, 149-161. Cortina, J.M. (1993). What is Coefficient Alpha: An Examination of Theory and Application. Journal of Applied Psychology, 78, 98-104. Costa, P. T., and McCrae, R.R. (1980). Enduring dispositions in adult males. Journal of Personality and Personal Psychology, 38(5), 793-800.

36

Dahlstrom,W. G.,Welsh, G., & Dahlstrom, L. (1972). An MMPI Handbook: Volume 1, Clinical Interpretations. Minneapolis, MN: University of Minnesota Press. Darwin, C. (1920). The Expression of Emotion in Man and Animals. New York, NY: Appleton-Century-Crofts. Ekman, P., & Davidson, R. J. (Eds.). (1994). The Nature of Emotions: Fundamental Questions. New York, NY: Oxford University Press. Emde, R. N. (1981). Changing Models of Infancy and the Nature of Early Development: Remodelling the Foundation. Journal of the American Psychoanalytical Association, 29, 179-219. Erikson, E. H. (1963).Childhood and Society .New York, NY: W.W. Norton & Society. Erikson, E. H. (1968). Identity: Youth and Crisis. New York, NY: W.W. Norton. Eysenck, H. J. (1960). The Structure of Human Personality. London, UK: Routledge and Kegan Paul. Fairbairn,W. R. D. (1952).An Object-Relations Theory of the Personality. New York, NY: Basic Books. Fischer, C. H. (1987). Assertivenss Training. In R. Corsinin (Ed.), Concise Encyclopedia of Psychology (pp. 91-92). New York: John Wiley & Sons. Freud, S. (1905). Three Essays on Sexuality. In J. Strachey (Ed.), The Standard Edition of the Complete Psychological Works of Sigmund Freud (Vol. 7.). London: The Hogarth Press and the Institute of Psychoanalysis. Goldfried, M. R., Stricker, G., & Weiner, I. (1971). Rorschach Handbook of Clinical and Research Applications. Englewood Cliffs, NJ: Prentice-Hall. Gouch, H. (1975). Manual for the California Psychological Inventory. Palo Alto, CA: Consulting Psychologists Press. Greenberg, J. R., & Mitchell, S. A. (1983). Object Relations in Psychoanalytic Theory. Cambridge, Mass: Harvard University Press. Guntrip, H. (1968). Schizoid Phenomena, Object-Relations and the Self. New York, NY: International Universities Press. Hathaway, S. R., & McKinley, J. C. (1943). Manual for the Minnesota Multiphasic Personality Inventory. New York, NY: Psychological Corporation. Heatherton, T. F., and Polivy, J. (1991). Development and Validation of a Scale for Measuring State Self-Esteem. Journal of Personality and Social Psychology, 60, 895-910.

37

Hogan, R. T., & Hogan, J. (1992). Hogan Personality Inventory Manual. Tulsa, OK: Hogan Assessment Systems. Hogan, R. T., Johnson, J.,& Briggs, R. (Eds.). (1997). Handbook of Personality Psychology. New York, NY: Morgan Kaufman. Husted, B. W. (1989). Trust in Business Relations: Directions for Empirical Research. Business and Professional Ethics Journal, 8, 23-40. Jöreskog, K.G., and Sörbom, D. (2000) LISREL 8.5 User’s Reference Guide. Chicago, IL: Scientific Software International. Jung, C. G. (1923). Psychological Types. New York, NY: Harcourt, Brace, Jovanovich. Kagan, J. (1989). Temperamental Contributions to Social Behavior. American Psychologist, 44, 668-674. Kernis, M. (1993). The Roles of Stability and Level of Self-Esteem in Psychological Functioning. In R. Baumeister (Ed.), Self-Esteem: The Puzzle of Low Self-Regard (pp. 167-182). New York: Plenum. Kets de Vries, M. (1993). Leaders, Fools, and Impostors: Essays of the Psychology of Leadership, San Francisco: Jossey-Bass. Kets de Vries, M. F. R. (1995), Life and Death in the Executive Fast Lane: Essays on Irrational Organizations and Their Leaders, San Francisco: Jossey-Bass. Kets de Vries, M.F. R. (2004a). Leadership Group Coaching in Action. INSEAD working paper. Kets de Vries, M. F. R. (2004b). The Global Executive Leadership Inventory: Participant Workbook. New York: John Wiley & Sons Kets de Vries, M. F. R. (2004c). The Global Executive Leadership Inventory: Facilitator’s Guide. New York: John Wiley & Sons Kets de Vries, M. F. R. (2005a). The Personality Audit: Facilitator’s Guide. Forthcoming. Kets de Vries, M. F. R. (2005b). The Personality Audit: Participant Workbook. Forthcoming. Kets de Vries, M., and Florent-Treacy, E. (1999). The New Global Leaders, San Francisco: Jossey-Bass. Kets de Vries, M., Shekshnia, S., Korotov, K., and Florent-Treacy, E. (2004). The New Russian Business Leaders, Cheltenham: Edward Elgar.

38

Kets de Vries, M., Vrignaud, P., and Florent-Treacy, E. (2004). The Global Leadership Life Inventory : Development and Psychometric Properties of a 360-degree Feedback Instrument. International Journal of Human Resource Management, 15 (3), 475-492. Kilberg, R. R. (2000). Executive Coaching. Washington, DC: American Psychological Association. Kohlberg, L. (1981). The Philosophy of Moral Development. New York, NY: Harper & Row. Kouzes, J., and Posner, B. (2001). Leadership Practices Inventory. San Francisco: Jossey-Bass. Lazarus, R. (1991). Emotion and Adaptation. New York, NY: Oxford University Press. Leary, M., and MacDonald, G. (2003). Individual Differences in Self-Esteem: A Review and Theoretical Integration. In M. Leary, and Tangney, J. (Ed.), Handbook of Self and Identity (pp. 401-418). New York: The Guilford Press. Lichtenberg, J. (1989). Psychoanalysis and Motivation. Hillsdale, NJ: Analytic Press. Lichtenberg, J. D. (1991). Psychoanalysis and Infant Research. New York, NY: Lawrence Erlbaum. Lichtenberg, J. D., Lackmann, F. M., & Forshage, J. L. (1992). Self and Motivational Systems: Toward a Theory of Psychoanalytic Technique. New York, NY: Analytic Press. Lichtenberg, J. D., & Schonbar, R.A. (1992). Motivation in Psychology and Psychoanalysis. In: J.W. Barron & M. N. Eagle & D. L.Wolitzky (Eds.), Interface of Psychoanalysis and Psychology. Washington, DC: American Psychological Association, 11-36. London, M. & Beatty, RW. (1993). 360-degree feedback as a competitive advantage. Human resource Management, 39, 353-372. London M., Smither, J, and Adsit, D. (1997), Accountability: The Achilles’ Heel of Multisource Feedback, Group and Organization Management, 22, 162-184 Lord, F., and Novick, M. (1968). Statistical Theories of Mental Test Scores. Oxford: Addison-Wesley Mahler, M. S., Pine, F., & Bergman, A. (1975). The Psychological Birth of the Human Infant. New York, NY: Basic Books. McNulty, J. (2000). Five Factor Model of Personality. In A. Kazdin (Ed.), Encyclopedia of Psychology (Vol. 3, pp. 375-376): American Psychological Association.

39

MacCallum, R. (1986). Specification searches in covariance structure modeling . Psychological Bulletin, 100 , 107-120 . MacCallum, R.C., Wegener, D.T., Uchino, B.N. & Fabrigar, L.R. (1993). The problem of equivalent models in applications of covariance structure analysis . Psychological Bulletin , 114,185-199 . McCrae, R. R., and Costa, P.T. (1987). Validation of the five-factor model of personality across instruments and observers. Journal of Personality and Social Psychology, 52, 81-90. Messick, S. (1994). Foundations of validity: Meaning and consequences in psychological assessment . European Journal of Psychological Assessement , 10, 1-9 . Millon, T. (1983). Millon Multiaxial Inventory Manual (third ed.). Minneapolis, MN: Interpretive Scoring Systems. Mohrman, A., Resnik-West, S., and Lawler, E. (1989). Designing Performance Appraisal Systems. San Francisco: Jossey-Bass. Murray,A. (1938). Explorations in Personality. New York, NY: Oxford University Press. Nunnally, J.C. (1978). Psychometric Theory. New York: McGraw-Hill. Peterson, R.A. (1994). Cronbach’s Alpha Coefficient: A Meta-analysis. Journal of Consumer Research, 21, 381-391 Piaget, J. (1952). The Origins of Intelligence in Children. New York, NY: International Universities Press. Pine, F. (1985). Developmental Theory and Clinical Process. New Haven, CT: Yale University Press. Plutchik, R. (1962). The Emotions: Facts, Theories and a New Model New York, NY: Random House Posner, B., and Kouzes, J. (1988). Development and Validation of the Leadership Practices Inventory, Educational and Psychological Measurement, 48: 483-496. Posner, B., and Kouzes, J. (1993). Psychometric Properties of the Leadership Practices Inventory - Updated, Educational and Psychological Measurement, 53: 191-199. Raftery, A. (1993). Bayesian model selection in structural equation model . in K.A.Bollen, J.S.Long (Eds) . Testing structural equation models . pp. 163-180 Newsbury Park CA: Sage Publications . Reich,W. (1933). Character Analysis. New York, NY: Farrar, Straus, and Giroux.

40

Ring, P. (1998). Trust. In N. Nicholson (Ed.), The Blackwell Encyclopedic Dictionary of Organizational Behavior (pp. 584-586). Malden, MA, and Oxford, UK: Blackwell Business. Shneidman, E. S., Joel, W.,& Little, K. B. (1951). Thematic Test Analysis. Orlando, FL: Grune & Stratton. Shrout, P. E., & Fleiss, J. L. (1979). Intraclass correlations: Uses in assessing rater reliability . Psychological Bulletin, 86, 420-428 . Spitz, R. A. (1965). The First Year of Life. New York, NY: International Universities Press. Stassen Berger, K. (1998). The Developing Person through the Life Span. New York, NY: Worth Publishers. Sullivan, H. S. (1953). The Interpersonal Theory of Psychiatry. New York, NY: Norton. Tomkins, S. S. (1962). Affect Imagery Consciousness: The PositiveAffects (Vol. I). New York, NY: Springer Publishing Company. Tomkins, S. S. (1963). Affect Imagery: Consciousness: The Negative Affects (Vol. II). New York, NY: Springer Publishing Company. Tomkins, S. S. (1991). Affect Imagery Consciousness: The Negative Affects, Anger and Fear (Vol. III). New York, NY: Springer Publishing Company. Tomkins, S. S. (1992). Affect Imagery Consciousness: Cognition, Duplication and Transformation of Information (Vol. IV). New York, NY: Springer Publishing Company. Tomkins, S. S. (1995). Script Theory: Differential Magnification of Affect. In V. Demos (Ed.), Exploring Affect: Selections from the Writings of Silvan S. Tomkins, New York, NY: Cambridge University Press, 312-410. Twenge, J. (2001). Changes in Women's Assertiveness in Response to Status and Roles: A Cross-Temporal Meta-Analysis, 1931-1993. Journal of Personality and Social Psychology, 81(1), 133-145. Watson, D., Clark, A., & Tellegen, A. (1988). Development and Validation of Brief Measures of Positive and Negative Affect: The PANAS Scales. Journal of Personality and Social Psychology, 54: 1063-1070 Westen, D., and Heim, A.K. (2003). Disturbances of Self and Identity in Personality Disorders. In M. Leary, and Tangney, J. (Ed.), Handbook of Self and Identity. New York: The Guilford Press. .

41

42

White, R. (1959). Motivation Reconsidered: The Concept of Competence. Psychological Review, 66, 297-333. White, R. (1966). Lives in Progress. New York, NY: Holt, Rinehart and Winston. Winnicott, D.W. (1975). Through Pediatrics to Psycho-Analysis. New York, NY: Basic Books. Yammarino, F., & Atwater, L.E. (1993). Understanding Self-Perception Accuracy: Implications for Human Resource Management. Human Resource Management, 32, 2331-2347 Yammarino, F.& Atwater, L.E. (1997). Do managers see themselves as others see them ? Organizational Dynamics, spring, 35-44.